- Challenge
- With the increasing adoption of gPlug products, the volume of support requests increased significantly. Many requests were repetitive or referred to information already available in previous emails, documentation, or on the website. At the same time, this knowledge existed in an unstructured form within email inboxes and could only be utilised with considerable manual effort. This led to inefficient processes and extended response times in support operations.
- Solution
- We developed a fully Swiss-operated AI solution based on Retrieval Augmented Generation (RAG), which automatically unlocks and makes historical support knowledge usable. Support emails are automatically extracted, processed by a Large Language Model, cleaned of personal data and irrelevant content, and transformed into structured question-and-answer pairs. From this information, compact knowledge snippets are generated and stored in a vector database together with content from technical documentation and the company website. Through Open WebUI, employees gain access to various preconfigured AI agents, including a Q&A assistant for rapidly answering support questions and an email agent for creating response drafts. All model inference is performed using open-source LLMs hosted by a Swiss provider, ensuring data protection, data sovereignty, and information security at all times.
- Outcome
- With the introduction of the solution, gPlug can, for the first time, make the knowledge previously distributed across support emails centrally available, structured, and usable across the entire organisation. As a result, support processes become significantly more efficient: responses can be created faster and more consistently, existing knowledge can be reused across channels, and the quality of information provided is sustainably improved through context-based AI support. Standard inquiries can be processed semi-automatically, allowing support employees to focus more on complex issues and value-adding tasks. At the same time, the fully Swiss-operated solution ensures maximum data sovereignty, compliance with all data protection requirements, and complete independence from foreign legislation such as the US Cloud Act.
